Although analyses of DNA, RNA, and protein expression can elucidate the phenotype of a cell, ultimately it is the interactions between proteins that determine metabolic function. Understanding protein-protein interactions is vital to the study of cancer, and the aim of most new chemotherapeutic drugs is to disrupt aberrant interactions. Recent advances in the quantification of tumor biomarkers in patient samples show great promise in predicting patient outcome and response to treatment. However, there is no good way to assess the co-localization of proteins in tissue samples. The successful evolution of bio-specific therapies and associated pharmaco-diagnostics may ultimately require quantitative measures of protein-protein interactions within each patient's tumor. From the basic science perspective, new methods in co-localizing proteins in tumor samples will advance the understanding of carcinogenesis and the subsequent development of targeted therapies. We hypothesize that engineering robust methods for assessing protein-protein interactions in human tumor samples will significantly improve the analysis of patient specimens and ultimately speed the development of targeted patient-specific therapies. Protein interactions are the gears which drive cells. In tumor cells, many of these processes are corrupted. Today, we can study the expression of individual proteins in patient tumor samples; however, we cannot look at the interactions between these proteins in patient tumors. Our goal is to create new methods for recognizing and studying such interactions. We hope that these methods will improve our understanding of cancer and provide means to assess which tumors will respond to specific therapies. [unreadable] [unreadable] [unreadable]